[PATCH v2 5/8] dm: Introduce DMA constraints into the core device model

[Simon Glass]

Hi Nicolas,

On Thu, 10 Dec 2020 at 04:43, Nicolas Saenz Julienne
<nsaenzjulienne at suse.de> wrote:
>
> Calculating the DMA offset between a bus address space and CPU's every
> time we call phys_to_bus() and bus_to_phys() isn't ideal performance
> wise, as it implies traversing the device tree from the device's node up
> to the root. Since this information is static and available before the
> device's initialization, parse it before the probe call an provide the
> DMA offset in 'struct udevice' for the address translation code to use
> it.
>
> Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne at suse.de>
>
> ---
> Changes since v1:
>  - Update commit message so as to explain better the reasoning behind
>    this
>
>  drivers/core/device.c | 24 ++++++++++++++++++++++++
>  include/dm/device.h   |  1 +
>  2 files changed, 25 insertions(+)
>

Firstly, please add a sandbox test for this. All driver model code is
tested this way.

> diff --git a/drivers/core/device.c b/drivers/core/device.c
> index 4b3dcb3b37..4255bea24d 100644
> --- a/drivers/core/device.c
> +++ b/drivers/core/device.c
> @@ -421,6 +421,28 @@ fail:
>         return ret;
>  }
>
> +void device_get_dma_constraints(struct udevice *dev)
> +{
> +       phys_addr_t cpu;
> +       dma_addr_t bus;
> +       u64 size;
> +       int ret;
> +
> +       if (!dev_of_valid(dev))
> +               return;
> +
> +       ret = dev_get_dma_range(dev, &cpu, &bus, &size);
> +       if (ret) {
> +               /* Don't complain if no 'dma-ranges' were found */
> +               if (ret != -ENODEV)

Don't use -ENODEV, that is very confusing. Maybe -ENOENT

> +                       dm_warn("%s: failed to get DMA range, %d\n",
> +                               dev->name, ret);

Needs to return an error

> +               return;
> +       }
> +
> +       dev->dma_offset = cpu - bus;
> +}
> +
>  int device_probe(struct udevice *dev)
>  {
>         const struct driver *drv;
> @@ -482,6 +504,8 @@ int device_probe(struct udevice *dev)
>                         goto fail;
>         }
>
> +       device_get_dma_constraints(dev);

Check error and fail

Also add a new CONFIG and use CONFIG_IS_ENABLED() to check whether to
use this feature. Some boards won't need it.

> +
>         ret = uclass_pre_probe_device(dev);
>         if (ret)
>                 goto fail;
> diff --git a/include/dm/device.h b/include/dm/device.h
> index 5bef484247..59f711e3dd 100644
> --- a/include/dm/device.h
> +++ b/include/dm/device.h
> @@ -161,6 +161,7 @@ struct udevice {
>  #ifdef CONFIG_DEVRES
>         struct list_head devres_head;
>  #endif
> +       u64 dma_offset;

ulong? Also needs a comment and needs to be conditional on CONFIG

>  };
>
>  /* Maximum sequence number supported */
> --
> 2.29.2
>

Regards,
Simon